13 January 2020
This is a copy of an email I sent to the Lagukan mailing list.
Hey everyone. It's been a while since I sent out an update about Lagukan. I've made a lot of strategy changes which I'll describe below, but the short version is that I haven't been doing feature development on Lagukan since November; instead, I've been working on a new product (a general-purpose recommender system, which can still be used to discover music but won't integrate directly with a music player). After I launch it—hopefully within a week or two—I'm planning to come back and polish off Lagukan. That'll include bug fixes, better offline support in Android, algorithm improvements, possibly a Spotify Connect web client (so you can use Lagukan + Spotify from a web browser without installing anything), perhaps some enhancements to the desktop app, and maybe still an iPhone app. I have to be realistic about doing an iPhone app though, since I don't know if I'll for sure be able to justify the time it would take.
Findka, the new product
After I sent out the last email update in November, I started thinking more seriously about business strategy/how to make money. I also had a small existential crisis—the synopsis being that I thought maybe Spotify's recommendation algorithm actually was just fine, I just hadn't been using it correctly. (If true, that would've removed most of Lagukan's value proposition). My fears turned out to be unfounded. I tried using Spotify quite a lot over the past couple months, but it never got any smarter even with all the data I was giving it. In contrast, I've been enjoying Lagukan immensely :).
However, that experience did help me think sincerely about what the alternatives were. In August, I realized there was a good opportunity to build a general-purpose recommender system (which I've called "Findka"). Users could import their usage data from all the apps they use (currently, the prototype supports Spotify, Goodreads, RSS and YouTube) and then we can correlate all the data and make recommendations across content types. E.g. we could recommend movies based on what books you like.
This also has the advantage of being much, much easier to bootstrap than Lagukan, since it doesn't require getting rights to play music. In addition, domain-specific recommender systems like Lagukan can be added on as complementary products. If you connect your Lagukan account to Findka, then they could share data etc. Plus, if I build a user base with Findka first, I can use that to help promote (especially to artists) a new, independent streaming service for Lagukan.
I had been thinking since August that I wanted to get Lagukan going and then start work on Findka. But in November, I decided that it made a lot more sense to switch that around and do Findka first. So that's what I've been working on.
Feedback and Updates